CFD simulation and aerodynamic loads in RWIND

Students & Education
, , ,
1

Dear Sir or Madam,

My name is Amirhossein Rashidi, and I am a Master's student in the Civil Engineering program with a specialization in Structural Engineering at the University of Duisburg-Essen.

I have a technical question regarding wind load determination in the RWIND program. Since my Master's thesis deals with the CFD-based determination of aerodynamic loads, I would greatly appreciate your expert assessment.

As part of my Master's thesis, I am investigating the aerodynamic loading of a wind turbine using various CFD programs. The aim of the work is to compare the surface pressures resulting from a CFD simulation with the wind loads according to Eurocode and then to contrast these in an FEM calculation. I want to examine to what extent the realistic pressure distributions from the CFD simulation deviate from the normative wind loads specified by Eurocode.

Before analyzing the actual wind turbine, I initially created a simplified reference model. This consists of a wall with a circular opening in the middle. With this model, I want to demonstrate that the wind loads calculated according to Eurocode are significantly higher than the pressure distribution that results under realistic flow conditions in a CFD simulation.

For this purpose, I simulated a wind tunnel with the model located in the center. After completion of the CFD simulation, the pressure distribution on the surface was exported in order to apply it as surface loads in an FEM program. Finally, these loads are to be compared with the wind loads calculated according to Eurocode.

However, I encountered a problem: The surface pressures resulting from the CFD simulation are about four to five times lower than the wind loads calculated according to Eurocode for the same model. This significant discrepancy unsettles me, which is why I would like to receive your assessment.

My calculation approach is as follows:

For the CFD simulation, the inlet velocity was defined based on the mean wind velocity (v_m(z)) according to Eurocode. Additionally, turbulence was considered using the turbulence intensity (I_v(z)) in ANSYS Fluent. The pressure distribution on the structure was then determined from the simulation.

For the FEM calculation, on the other hand, the wind loads are determined based on the peak velocity pressure (q_p(z)) according to Eurocode. The design load is given by

wā‚‘ = cā‚šā‚‘ Ā· qā‚š(z)

and is then applied as a uniformly distributed surface load on the model.

As I understand it, the loads obtained in this way should basically be comparable with the pressure values from the CFD simulation. However, the results differ significantly.

My main question is therefore:

Is my methodological approach fundamentally correct, especially regarding the choice of inlet velocity for the CFD simulation? Should the mean wind velocity (v_m(z)) actually be used as the wind velocity, while the structural verification works with the peak velocity pressure (q_p(z))? Or is there possibly a conceptual error here that explains the large difference between the CFD results and the Eurocode values?

I would greatly appreciate it if you could briefly assess my approach or provide me with a hint about the cause of this significant deviation.

Thank you very much in advance for your time and support.

Best regards
Amirhossein Rashidi

2

Dear Mr. Rashidi,

Thank you very much for your thoughtful question. This is a very interesting topic and highly relevant for understanding the relationship between code-based wind loading and CFD simulations.

We have published several validation examples comparing RWIND results with Eurocode-based wind load calculations, which may be helpful for your research. I would recommend taking a look at these examples, as they provide useful insights into the similarities and differences between the two approaches.

:backhand_index_pointing_right: https://www.dlubal.com/en/downloads-and-information/examples-and-tutorials/verification-examples/009051

:backhand_index_pointing_right: https://www.dlubal.com/en/downloads-and-information/examples-and-tutorials/verification-examples/009048

:backhand_index_pointing_right: https://www.dlubal.com/en/downloads-and-information/examples-and-tutorials/verification-examples?q="Mahyar%20Kazemian,%20M.Sc."&f=_

From my perspective, it is generally expected that Eurocode produces more conservative wind loads than CFD. This is because the Eurocode design methodology is intentionally conservative and accounts for several aspects that are not necessarily represented in a standard steady-state CFD simulation, including:

  • Peak velocity pressure rather than mean wind velocity,

  • Unfavorable wind directions and statistical extremes,

  • Safety and reliability considerations inherent in the code calibration,

  • Conservative pressure coefficients derived from experimental databases,

Therefore, under comparable conditions, I would generally expect Eurocode wind loads to be somewhat higher than CFD results. However, a difference of approximately four to five times is larger than what would normally be expected and suggests that the comparison methodology should be examined carefully.

One important aspect is the definition of the inlet boundary conditions. If the CFD simulation uses the mean wind velocity profile, while the Eurocode loads are based on the peak velocity pressure, the two approaches are not directly comparable. A more representative comparison can be achieved by defining the inlet boundary conditions using a wind profile corresponding to the peak wind conditions while simultaneously incorporating the appropriate atmospheric turbulence intensity. This approach allows the CFD simulation to represent the atmospheric boundary layer more consistently with the assumptions underlying the Eurocode formulation.

An example of such an inlet definition is shown in the attached reference, where both the velocity profile and turbulence intensity are specified consistently. Using this methodology should lead to CFD results that are more comparable to Eurocode predictions.

Finally, it is also important to verify that the same reference height, terrain category, air density, pressure definition (static vs. pressure coefficient), and pressure averaging procedure are used in both approaches, as these factors can also contribute to noticeable differences.

I hope these comments are helpful. Please feel free to contact us if you have further questions, and I wish you every success with your Master's thesis.

Kind regards,

Mahyar Kazemian

1 Like